Papers summary

• ClabbersGrueneAbrahams2017 : **Clabbers, M. T. B., Genderen, E. Van, Wan, W., Wiegers, E. L., & Gruene, T. (2017). Protein structure determination by electron diffraction using a single three-dimensional nanocrystal. Acta Crystallographica Section D, 73, 738–748. doi
• LatychevskaiaAbrahams2019 : Latychevskaia, T., & Abrahams, J. P. (2019). Inelastic scattering and solvent scattering reduce dynamical diffraction in biological crystals. Acta Crystallographica Section B: Structural Science, Crystal Engineering and Materials, 75, 523–531. doi
• SubrimanianSpence2015 : Latychevskaia, T., & Abrahams, J. P. (2019). Inelastic scattering and solvent scattering reduce dynamical diffraction in biological crystals. Acta Crystallographica Section B: Structural Science, Crystal Engineering and Materials, 75, 523–531. doi
• CowleyMoodie1957 : Cowley, J. M., & Moodie, A. F. (1957). The scattering of electrons by atoms and crystals. I. A new theoretical approach. Acta Crystallographica, 10(10), 609–619. doi This is the original paper on the multislice method.

• oleynikov2007 : Oleynikov, P., Hovmöller, S., & Zou, X. D. (2007). Precession electron diffraction: Observed and calculated intensities. Ultramicroscopy, 107(6–7), 523–533. doi

• crystal bonding

ClabbersGrueneAbrahams2017

Protein structure determination by ED of 3D nanocrystal(Experimental) : dimeric polymorph hen egg-white lyzozyme (Diffracting volume $0.14\mu m^3$) using Molecular Replacement from monomeric polylalanine model and standard X-ray protein crystallography software.

The claim of a successful solution despite subramanianspence2015 is due to lack of inelastic modelling in multislice simulation. This is backed up by the fact that Friedel pair symmetry is not more violated than for standard X-ray data.

Software/Method	description
Data acquisition
TEM	Talos Arctica/Titan Krios, Parallel beam rotation $200keV$
detector	$1024\times 1024$ Timepix pixel($0.05\times0.05 mm$ each)
Data processing	XDS
Structure solution
XSCALE	scaling
POINTLESS	MTZ format conversion
AIMLESS	merging
TRUNCATE	Structure factor amplitudes
CHAINSAW	model creation
PHASER	Molecular replacement
Buccaneer/REFMAC5	Side chains placement
COOT	fitting missing residues
Refinement
REFMAC	EXPDTA ELECTRON CRYSTALLOGRAPHY, SOURCE ELECTRON MB flags
SFTOOLS	non measured observation removal
FREERFLAG	unique test sets creation

Micrograph (0.2x0.5x1.4nm)	Diffraction data	Fo vs Fc

LatychevskaiaAbrahams2019

Inelastic scattering and solvent scattering significantly mitigate the effect of dynamical diffraction.

Probabilities of scattering events	Inelastic fraction of dynamical scattering	Friedel symmetry violation solvent scattering

SubramanianSpence2015

Strong dynamical diffraction prevents single scattering approximation based techniques at crystal thickness above 100-200nm.

Multislice details (ZMULT package) :

• lysozyme Tetragonal $P4_32_12$ space group ( here or here), a=b=79.1A, c=37.8A.
• Structure factor from experimentally refined RHF Doyle-Turner data (H not included)
• 90 beams at 200keV

Friedel symmetry violation $I_{-440}$	Omit density map : threshold error 37%	MR distinguishable error limit 34%	Thickness limit $T_{R_f=0.3}\approx 100nm$

Oleynikov2007

Comparison of multislice and exp for precessionED and SAED.

figures	description
	Ewald radii during precED. $R_{n,in/out}=\sqrt{K^2-(K_z-nD)^2}\mp K\sin\varphi$ are intersections with ZOLZ and HOLZ.
	Integration path of the excitation error during precession ED where $s_g=g^2/2K-g_{xy}\varphi\cos(\alpha-\omega)-g_z$ with $\omega$ azimuthal angle. Oscill is the kinematic rocking curve function $t^2 sinc^2(ts_g)/\xi_g$ where $\xi_g=K/V_g$
	integration correction factor $A_{corr}$ where $I_g^{prec}=I\int_o^{2\pi}oscill(s_g(\omega))d\omega=IA_{corr}$ for h=1..30
	Reflections for azim=0 with forbidden reflections h00
	Reflections for azim=90 with forbidden reflections 0k0
	Rfactor for precession ED and SAED run with multislice compared with experimental data in the [001] direction for $Cs_{0.44}Nb_{2.54}W_{2.46}O_{14}$. precED simus were run for 28 patterns with thickness in range(60,600,20) each of them made of 250 diffraction patterns.